NaN values throwing different results using naniar

[manisharora]

Can someone explain the below results. It's really weird how NaN is detected:

[Snehal_Patel]

The NaN and Inf in the second vector are implicitly coerced into characters because of other character elements in the vector ("." and "missing"), whereas the NA is maintained. "Not a Number" (NaN) and "infinity" (Inf) do not makes sense to keep in a character vector as "special values" because they have no meaning in a character vector (as they would in numeric vectors). In contrast, the NA (i.e., missing value) still makes sense to keep in a character vector.

x = c(NA, NaN)
x
[1] NA NaN
class(x)
[1] "numeric"
are_na(x)
[1] TRUE TRUE
x = c(NA, NaN, Inf, ".", "missing")
x
[1] NA "NaN" "Inf" "." "missing"
class(x)
[1] "character"
are_na(x)
[1] TRUE FALSE FALSE FALSE FALSE

Hope that makes sense,
Snehal

[manisharora]

Thanks, yeah that definitely does make sense.